Abstract
Objective
Antibiotic overuse leading to antimicrobial resistance is a global public health concern. Clinical trials have demonstrated that procalcitonin-based decision-making for antibiotic therapy can safely decrease inappropriate antibiotic use in patients with respiratory infections and sepsis, but real-world data are scarce. This study sought to assess the impact of a procalcitonin-based antibiotic stewardship program (protocol plus education) on antibiotic use in community hospitals.
Methods
An observational, retrospective, matched cohort study was conducted. Eligible patients treated in hospitals with a procalcitonin-based protocol plus education (Procalcitonin cohort hospitals) were matched to patients admitted to facilities without procalcitonin testing (Control cohort hospitals) using a 1:2 ratio. The Control hospitals were facilities where procalcitonin testing was not available on site. Patient matching was based on: (1) age, (2) gender, (3) admission diagnosis code using groupings of the International Classification of Diseases, 10th Revision, (4) whether patients were admitted to the intensive care unit, and (5) whether a blood culture test was performed. Procalcitonin cohort hospitals implemented a quality improvement initiative, where procalcitonin was available, used regularly, and clinicians (physicians and pharmacists) were educated on its use.
Results
After adjustment, patients in the Procalcitonin cohort had 1.47 fewer antibiotic days (9.1 vs. 8.5 days, 95%CI: −2.72; −0.22, p = .021). There was no difference in length of stay or adverse clinical outcomes except for increase in acute kidney injury (odds ratio = 1.26, 95%CI: 1.01; 1.58, p = .038).
Conclusions
Patients with respiratory infections and sepsis in hospitals utilizing a procalcitonin-based protocol coupled with education received fewer days of antibiotic therapy.
Introduction
Antibiotic (AB) misuse and overuse has a significant impact on the local and global ecology of AB resistance, which represents a rising global public health problem. In 2013, the Centers for Disease Control and Prevention (CDC) estimated that in the United States (US), more than two million people were affected by AB-resistant infections annually, resulting in at least 23,000 deaths with an estimated cost of up to $20 billionCitation1. In addition to the threat of AB resistance, overuse of AB can lead to adverse effects such as renal dysfunction and Clostridioides difficile infectionsCitation2–5.
Due to concerns about rising rates of AB resistance, the US government released the “National Action Plan for Combating Antibiotic Resistant Bacteria” recommending that all healthcare facilities implement AB stewardship programs (ASPs)Citation6,Citation7. Effective ASPs, with appropriate drug product selection, dosing, route of administration, and duration of antimicrobial therapy, in conjunction with a comprehensive infection control program, have been shown to limit the emergence and transmission of AB-resistant microorganismsCitation8.
Biomarker-guided AB decision-making, which may identify patients for whom AB are not needed or can be safely stopped, is an example of a stewardship intervention that can improve AB use. Procalcitonin (PCT) is one such biomarker. PCT-guided treatment decisions for initiation or discontinuation of AB therapy can reduce AB use in patients with lower respiratory tract infections (LRTIs; i.e. acute bronchitis, exacerbations of chronic obstructive pulmonary disease [COPD], and pneumonia) and sepsis, without adverse effects on mortality, length of stay in the intensive care unit (ICU), or length of hospital stayCitation9–12. A recent randomized trial also found that PCT guidance may have benefits on long-term patient outcomes, including lower rates of infection-associated adverse events at 6 months among patients with sepsisCitation13. A 2018 systematic review and meta-analysis of 11 RCTs evaluating the impact of PCT guidance in patients with LRTIs found statistically significant reductions in AB duration (weighted mean difference: −2.15 days, 95% confidence interval [CI]: −3.30; −0.99) and in the odds of initiating AB (odds ratio: 0.26, 95%CI: 0.13; 0.52)Citation10. Similar reductions in AB use have been demonstrated in sepsis with decreased AB exposure (weighted mean difference: −1.49 days, 95%CI: −2.27; −0.71)Citation10. In both reviews, PCT guidance had no deleterious consequence on patient outcomes (lengths of hospital/ICU stays and mortality).
While numerous RCTs exist, evidence from real-world studies is limited. Furthermore, existing literature underrepresents real-world studies that involve community hospitals (CH) and critical-access hospitals (CAH). These facilities often lack access to experts in infectious diseases and AB stewardship and thus are searching for attainable methods to demonstrably improve AB use.
We evaluated the effects of PCT-guided AB therapy coupled with an education program as part of a quality improvement initiative. The primary objective was to assess if a PCT-based protocol (in-house PCT testing capacity, PCT results displayed in the patient’s electronic health record [EHR], and education) could reduce AB exposure (number of days of AB therapy) in patients with respiratory infections and sepsis in CH settings, compared with facilities with neither on-site PCT testing nor a protocol. Secondary objectives assessed differences in length of hospital stay, length of ICU stay, in-hospital mortality, readmission within 30 days, acute kidney injury (AKI) during the hospital stay, and hospital-onset Clostridioides difficile infection (HO-CDI).
Methods
Data source
De-identified patient data from EHRs of a large US CH system of more than 100 hospitals across 20 states were retrieved from 1 March 2018 to 17 September 2018.
Sample selection and construction
An observational, retrospective, matched cohort study was conducted. The study received approval from the hospital system’s institutional review board. Hospitals were categorized into “Treatment” and “Control” hospitals. “Treatment hospitals” met the following criteria: (1) had in-house PCT testing available, (2) had more than 10% of unique patients with PCT testing results during the study period, and (3) participated in the quality improvement intervention. The quality improvement intervention was multi-disciplinary and included a continuing education-accredited clinician lecture (live or recorded) that provided continuing medical education credit to infectious diseases physicians, emergency physicians, pulmonologists, hospitalists, internists, and intensivists, and Accreditation Council for Pharmacy Education – Continuing Education (ACPE-CE) credit for pharmacists. A clinical surveillance and decision support system (Sentri7)Footnotei that was already integrated into pharmacists’ workflow at all facilities sent real-time alerts to pharmacists at Treatment hospitals to intervene based on the PCT results or to order repeat testing based on serum PCT results (Supplemental material).
In facilities where pharmacists were not authorized by the Pharmacy and Therapeutics (P&T) committee to order PCT (2 out of 11 intervention facilities), recommendations were made to physicians to order repeat PCT testing based on an approved protocol (). If specific criteria were met (e.g. PCT below established threshold, white blood cell count normalized, no fever, active order for AB), the pharmacist was prompted to intervene with a recommendation for AB cessation. If the intervention was not accepted, the pharmacist was responsible for daily follow-up and repeat testing every 1–2 days until intervention acceptance. In the case of a positive PCT and no active order for AB, the pharmacist was notified and required to make an urgent intervention. In addition, the intervention included EHR integration where the PCT results included contextual recommendations (e.g. “Antibiotics STRONGLY DISCOURAGED if clinically stable,” “Antibiotics DISCOURAGED if clinically stable,” “Antibiotics ENCOURAGED,” “Antibiotics STRONGLY ENCOURAGED”), depending on the PCT value.
Figure 1. Algorithm Used for PCT Guidance among PCT Hospitals.
“Control hospitals” did not have PCT testing readily available. Facilities that performed PCT testing in a reference laboratory or neighboring facility were allowed in the Control hospitals as long as less than 10% of their eligible patients had PCT results during the study period. Facilities that were divested or closed during the study period and those without ICUs (e.g. orthopedic specialty hospitals, ambulatory surgery centers, women’s hospitals) were also excluded.
Patients admitted to a Treatment hospital formed the “PCT cohort,” and patients seen in a Control hospital formed the “Control cohort.” Cohort assignment was based on the facility, highlighting real-world changes due to PCT availability and PCT protocol compliance, rather than PCT use at the patient level. For the PCT cohort, enrollment occurred only after completion of all components of the quality improvement initiative.
For each patient, the first hospitalization during the study period meeting the following criteria was considered: (1) at least 18 years of age, (2) at least one AB administration during their hospital stay, (3) admitted through the facility’s emergency department, and 4) discharged with a discharge diagnosis related group (DRG) code for respiratory infection (DRG 177 − 179), COPD (DRG 190 − 192), bronchitis and asthma (DRG 202, 203), pneumonia (DRG 193 − 195), or sepsis (DRG 870 − 872). Exclusion criteria included: (1) admission to the hospital in the 30 days prior to the admission, and (2) patients in the Control cohort who received a PCT test during their hospital stay. The observation period was defined as the time from admission date until 30 days following discharge from the hospital.
Analysis
All eligible patients in the PCT cohort were matched to patients in the Control cohort using a 1:2 ratio. Patients were matched exactly on: (1) admission diagnosis code defined using groupings of the ICD-10, (2) sex, (3) age (±1 year), (4) admission to the ICU during the hospital stay, and (5) whether or not a blood culture test was performed. ICU admission and blood culture testing served as a proxy for illness severity.
The primary measured outcome was AB exposure during the hospital stay, measured in days of AB therapy (AB DOT). Each AB agent was counted separately; hence, a single calendar day with multiple agents was counted as multiple AB DOT. Secondary outcomes included hospital length of stay (in days), ICU length of stay (in days), 30-day readmissions, in-hospital mortality, onset of AKI during the hospital stay (per physician diagnosis and identified through ICD-10 codes), and HO-CDI.
Unadjusted differences between cohorts were calculated using the Wilcoxon rank-sum test for continuous outcomes and chi-square test for dichotomous outcomes. Adjusted differences were calculated using multivariable generalized estimating equations (PROC GENMOD in SAS) accounting for clustering at the hospital level. In addition to controlling for the matching criteria, analyses adjusted for hospital size, hospital affiliation with an academic institution, patient’s insurance type (Medicaid, self-pay/unknown), and indicators for black race or Hispanic ethnicity. Confidence intervals were generated using robust variance estimators and p-values were calculated using a two-sided α = 0.05. Adjusted differences (AD) were reported for continuous outcomes and adjusted odds ratios (OR) were reported for dichotomous outcomes.
All analyses were conducted using SAS 9.4.
Results
Forty hospitals from 14 states were considered for inclusion in the study. After exclusion of non-qualifying facilities, Treatment hospitals included 10 hospitals from 8 states, and Control hospitals included 28 hospitals from 10 states. Treatment hospitals were larger on average, with a mean bed size of 321, compared with 158 among Control hospitals. A higher proportion of Treatment hospitals than Control hospitals were affiliated with an academic institution (30% vs. 7%).
After matching, 2424 patients from Treatment hospitals were included in the PCT cohort and 4848 patients from Control hospitals were included in the Control cohort. Patients’ mean age at hospital admission was 68.1 (median = 70); 45.6% were male; 24.3% were admitted to the ICU during their hospitalization; and 85.6% had a blood culture performed (). PCT testing was conducted in 59.9% of patients in the PCT cohort; the mean number of tests per patient was 1.2 (median = 1). No patients in the Control cohort underwent PCT testing. The most common admission diagnosis codes (ICD-10) included “A41 – Other sepsis” (22.2%), “R06 – Abnormalities of breathing” (17.5%), “J18 – Pneumonia, unspecified organism” (12.9%), and “J44 – Other chronic obstructive pulmonary disease” (9.4%). An additional 24.1% of patients were admitted with other ICD-10 codes beginning with “R” (“Symptoms and signs involving the circulatory and respiratory systems”). All other diagnosis codes (14.1% of patients) were grouped together.
Table 1. Unadjusted differences and proportion of patients with blood culture testing before and after matching.
Primary outcome
The mean unadjusted number of AB DOT was 9.1 (standard deviation [SD] = 7.0, median = 7) in the PCT cohort and 8.5 (SD = 7.6, median = 6) in the Control cohort (unadjusted difference = 0.6 days, p < .001). While the mean number of days before adjustment was higher in the PCT cohort, after adjustment using multivariable generalized estimating equations, the AB DOT was 1.47 days shorter among patients in the PCT cohort compared to the Control cohort (95%CI: −2.72; −0.22, p = .021).
Secondary outcomes
After adjustment, no statistically significant effect of the PCT-based protocol was observed on hospital length of stay (AD = –0.14 days, 95%CI: −0.83; 0.55, p = .690), ICU length of stay (AD = 0.03 days, 95%CI: −0.13; 0.18, p = .740), and 30-day readmissions (OR = 0.99, 95%CI: 0.77; 1.28, p = .946). Patients in the PCT cohort had higher odds of developing AKI (OR = 1.26, 95%CI: 1.01; 1.58, p = .038) than patients in the Control cohort, but in-hospital mortality was not different between groups (OR = 1.26 95%CI: 0.78; 2.05, p = .341). The low incidence of HO-CDI in the sample after matching (PCT cohort, n = 12 [0.5%], Control cohort, n = 26 [0.5%]) precluded comparison of this outcome ().
Table 2. Multivariable outcomes results.
Discussion
The results suggest that implementation of a PCT-based protocol coupled with education is both possible and can effectively reduce AB exposure in patients with LRTI and sepsis in community hospital settings. It appears that this intervention was safe as there were no statistically significant differences in hospital length of stay, length of stay in the ICU, 30-day readmissions, or in-hospital mortality although an increase in AKI was observed in the PCT cohort.
Interestingly, prior to multivariable adjustment, the number of days of AB was significantly higher in the PCT cohort than in the Control cohort. Following adjustment, however, the opposite was observed. Covariates included in the multivariable regressions aimed at controlling for the patients’ underlying disease severity, and included whether an ICU stay was required and blood cultures taken. Since more severely ill patients may require more AB agents for longer durations, it was important to adjust for these factors. Patients in the PCT cohort were recruited from facilities that were generally larger and more frequently affiliated with an academic center. Larger hospitals, both public and private, have been found to have a higher case mix index (CMI), which is an indicator for disease severity. Similarly, teaching hospitals (hospitals with physicians “enrolled in an Accreditation Council for Graduate Medical Education training program”) have higher CMI than nonteaching facilitiesCitation14. Therefore, despite matching on admission diagnosis, it is possible that within each diagnostic category, patients in the PCT cohort were more severely ill, which may explain why the covariate adjustment reversed the apparent impact of PCT on number of days of AB therapy. It is also possible that these more severely ill patients may receive concomitant antibiotics as part of an appropriate treatment course, further increasing the AB DOT in the PCT cohort prior to statistical adjustment. PCT hospitals also had a higher percentage of consults with infectious diseases specialists, perhaps indicating more severe or resistant infections. Finally, when limiting the analysis to patients with a blood culture, the PCT cohort had a greater mean number of AB DOT (9.6 vs. 8.7, p < .001), further suggesting that the PCT cohort included more severely ill patients. To note, since multiple AB could be administered on a single day, the higher AB DOT observed in the PCT cohort prior to adjustment did not necessarily entail a longer course of treatment in calendar days.
The impact of the admission diagnosis on the difference in AB DOT could be observed with the coefficients obtained from the multivariate adjustment. Diagnoses “R06 Abnormalities of breathing” and “J44 Other COPD” were both statistically significantly associated with a shorter AB DOT, while “J18 Pneumonia, unspecified organism” was associated with a longer AB DOT. This shows that the impact of PCT guidance in the real world is likely to change based on the patient’s condition.
Compared with previous RCTs of PCT in LRTI and sepsisCitation10, the effect size on AB duration was smaller. RCTs evaluate outcomes in highly-regulated environments and typically have dedicated resources to ensure improved compliance with PCT recommendations. Previous studies of PCT use in the real world have shown significant alterations in AB use, although the impact was less than that found in RCTsCitation15. This may be due to use of PCT in patient populations not typically included in RCTsCitation15. The effect size may have also been impacted by the heterogeneity of available services, patient mix, and general AB use between PCT hospitals and Control hospitals.
Previous trials of “real life” PCT use have been mixed in their impact. A recent RCT from Huang et al. found that PCT-guided use of AB did not result in fewer days of therapy compared with usual care (PCT group: 4.2 days, usual care: 4.3 days, p = .87)Citation16. Patients enrolled in that study had fewer comorbidities and less severe illness than many of the previous RCTs of PCT useCitation17. Additionally, AB therapy in the usual care group was commonly withheld when unnecessary and decisions taken at the sites (i.e. emergency departments, hospitals) were often overruled in the outpatient settingCitation16. The primary outcome measure in the Huang et al.Citation16 study was also different: they measured days of AB exposure while our study measured AB DOT, which takes into account the use of more than one agent on a given day.
Compared with some other previous real-world studiesCitation18,Citation19, the present study shows that PCT has a greater effect on reducing AB DOT. A 2017 retrospective cohort study of 20,750 critically ill patients (approximately 20% of patients with sepsis in US nonfederal hospitals) found that PCT use was associated with increased AB duration (adjusted relative risk: 1.17, 95%CI: 1.15; 1.18)Citation19. The authors observed that there were wide variations among hospitals’ and clinicians’ use of PCT in the community, leading to the conclusion that the lack of measureable benefits was the result of “[poor] implementation into real-world practice.”Citation19 In addition, the study was limited to the subset of patients admitted to ICUs. Hence, the results may not be generalizable to patients with less severe conditions. Our positive outcome may have been due to the method of implementation of PCT which included an educational intervention and included pharmacist accountability for PCT protocol follow-up.
There are other real-world studies that are consistent with our results. A study evaluating the impact of education on use of a PCT algorithm and stewardship practices to encourage adherence to the algorithm in a CH demonstrated a reduction in AB duration and in-hospital mortalityCitation20. Another study in hospitalized patients with uncomplicated pneumonia found that following clinician education on use of a PCT algorithm, mean duration of AB therapy and hospital length of stay was significantly shorterCitation21. Finally, Albrich et al. published their experience with “real life” PCT implementation in patients with LRTI finding that AB use was significantly decreased compared to previous RCT control groups without any signal for patient harmCitation15.
The proportion of patients with AKI was higher in the PCT cohort, and remained higher after multivariable adjustment. In patients with AKI, those in the PCT cohort had more AB DOT (10.7 vs. 9.5, p = .002). It is therefore possible that patients in the PCT cohort were more severely ill with a higher likelihood of developing AKI from use of nephrotoxic medications or interventions. Previous studies have found that intensive AB therapy leads to emergence or prolongation of renal failure in intensive care patientsCitation2,Citation22. Thus, by having more days of AB DOT, these patients may have been more at risk for AKI. Similarly, while not statistically significant, the observed OR greater than one for in-hospital mortality may reflect that the PCT cohort had more severe disease.
There are several limitations to our study. We attempted to mitigate differences between the PCT cohort and the Control cohort using matching and multivariate modeling, but there likely remained unobserved and/or unavailable confounders between cohorts, most notably in terms of comorbidities at admission, type and severity of illness, and nature of clinicians caring for the patients. This may have confounded the estimation of the benefit of PCT protocolization on AB use or the association between PCT guidance and AKI. Future studies should consider the use of a severity score (e.g. APACHE score) or biomarkers (e.g. white blood cell counts or C-reactive protein) as a means to account for variations in patients’ severity.
Furthermore, all data were retrieved from EMR between March and September of 2018. Results reported in this study are thereby only representative of the spring and summer of that year and may have been different during a different period.
There was variability in the implementation and acceptance of the educational interventions by participating clinicians. Sub-optimal attendance and understanding of the intervention may have blunted the impact of protocolized PCT. Pharmacists at PCT facilities who were part of the hospital-system stewardship team were utilized for PCT implementation to overcome these knowledge deficits in front-line prescribers and attempt to ensure compliance. Furthermore, compliance with the algorithm in the PCT cohort was not measured, which is known to be associated with the efficacy of PCT to safely reduce AB use in the real worldCitation15. However, compliance with PCT guidance is expected to vary in the real world. Patients’ clinical presentation and physician experience also contribute to the AB decision-making process. It is possible that the recommendation was not followed for unobserved reasons. Therefore, results of the current study should be interpreted as the real-world effectiveness of PCT guidance, rather than its clinical efficacy.
The current study was also performed using EHR from CHs in the US with support from a third-party electronic clinical surveillance and decision support program, and results may not apply to other settings, such as outpatient clinics.
Finally, the current study only focused on AB DOT as a measure of effectiveness. Dosing de-escalation of AB following PCT guidance would have been interesting to measure, albeit it was not assessed in the current study. Future research may consider AB de-escalation as an additional relevant outcome when evaluating the effectiveness of PCT guidance.
Conclusion
Implementation of a PCT-based protocol with on-site testing, education, and pharmacist support was associated with a reduction in days of AB therapy in patients with respiratory infections and sepsis in real-world practice among US CH.
Transparency
Declaration of funding
This work was supported by bioMérieux, Inc. bioMérieux is the manufacturer of the VIDAS family of instruments for the determination of procalcitonin in human serum or plasma (lithium heparinate) using the ELFA (Enzyme-Linked Fluorescent Assay) technique. B·R·A·H·M·S PCTFootnoteii is the property of Thermo Fisher. The study sponsor was involved in the study design (definition of the research objective and identification of the data source), the interpretation of data (review and interpretation of initial results), the review and approval of the manuscript, and the decision to submit the manuscript for publication.
Declaration of financial/other relationships
KD and SI are former employees of CHSPSC, LLC. PS has received research funding from bioMérieux SA, bioMérieux, Inc., Thermofisher, Roche, Siemens and Abbott. PT-L, NK, HCC, and SO are employees of Analysis Group, Inc., a consultancy that has received research funding from bioMérieux for participation in this study. TV has received research funding from Merck and Rebiotix and served as a consultant for bioMérieux. LZ is an employee of bioMérieux, Inc. Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.
Author contributions
All authors agree to be accountable for all aspects of the work.
Supplemental Material
Download MS Word (1.5 MB)Acknowledgements
The authors would like to thank Yi Zhong for her assistance with data analysis.
Notes
i Sentri7 is a registered trademark of Wolters Kluwer Health, Madison, WI, USA.
ii B·R·A·H·M·S PCT is a trademark of Thermo Fisher Scientific Inc., Waltham, MA, USA.
References
- Center for Disease Control and Prevention. Antibiotic resistance threats in the United States. U.S. Washington, D.C.: Department of Human and Health Services; 2020.
- Jensen JU, Hein L, Lundgren B, et al. Kidney failure related to broad-spectrum antibiotics in critically ill patients: secondary end point results from a 1200 patient randomised trial. BMJ Open. 2012;2(2):e000635.
- Stevens V, Dumyati G, Fine LS, et al. Cumulative antibiotic exposures over time and the risk of Clostridium difficile infection. Clin Infect Dis. 2011;53(1):42–48.
- Antharam VC, Li EC, Ishmael A, et al. Intestinal dysbiosis and depletion of butyrogenic bacteria in Clostridium difficile infection and nosocomial diarrhea. J Clin Microbiol. 2013;51(9):2884–2892.
- Brown KA, Khanafer N, Daneman N, et al. Meta-analysis of antibiotics and the risk of community-associated Clostridium difficile infection. Antimicrob Agents Chemother. 2013;57(5):2326–2332.
- Fridkin S, Baggs J, Fagan R, et al. Vital signs: improving antibiotic use among hospitalized patients. MMWR Morb Mortal Wkly Rep. 2014;63(9):194–200.
- Interagency Task Force for Combating Antibiotic-Resistant Bacteria. National action plan for combating antibiotic resistant bacteria. In: The white house, editor. Washington, DC: Interagency Task Force for Combating Antibiotic-Resistant Bacteria; 2015. p. 62.
- Hohn A, Heising B, Schütte J-K, et al. Procalcitonin-guided antibiotic treatment in critically ill patients. Langenbecks Arch Surg. 2017;402(1):1–13.
- Hey J, Thompson-Leduc P, Kirson NY, et al. Procalcitonin guidance in patients with lower respiratory tract infections: a systematic review and meta-analysis. Clin Chem Lab Med. 2018;56(8):1200–1209.
- Claxton A, Iankova I, Thompson-Leduc P, et al. Efficacy and safety of procalcitonin guidance in patients with suspected or confirmed sepsis: a systematic review and meta-analysis*. Crit Care Med. 2018;46(9):1560–1698.
- Bouadma L, Luyt C-E, Tubach F, et al. Use of procalcitonin to reduce patients’ exposure to antibiotics in intensive care units (PRORATA trial): a multicentre randomised controlled trial. The Lancet. 2010;375(9713):463–474.
- Meisner M. Update on procalcitonin measurements. Ann Lab Med. 2014;34(4):263–273.
- Kyriazopoulou E, Liaskou-Antoniou L, Adamis G, et al. Procalcitonin to reduce long-term infection-associated adverse events in sepsis. a randomized trial. Am J Respir Crit Care Med. 2021;203(2):202–210.
- Mendez CM, Harrington DW, Christenson P, et al. Impact of hospital variables on case mix index as a marker of disease severity. Popul Health Manag. 2014;17(1):28–34.
- Albrich WC, Dusemund F, Bucher B, et al. Effectiveness and safety of procalcitonin-guided antibiotic therapy in lower respiratory tract infections in “real life”: an international, multicenter poststudy survey (ProREAL). Arch Intern Med. 2012;172(9):715–722.
- Huang DT, Yealy DM, Filbin MR, et al. Procalcitonin-guided use of antibiotics for lower respiratory tract infection. N Engl J Med. 2018;379(3):236–249.
- Schuetz P, Christ-Crain M, Thomann R, et al. Effect of procalcitonin-based guidelines vs standard guidelines on antibiotic use in lower respiratory tract infections: the ProHOSP randomized controlled trial. Jama. 2009;302(10):1059–1066.
- Balk RA, Kadri SS, Cao Z, et al. Effect of procalcitonin testing on health-care utilization and costs in critically ill patients in the United States. Chest. 2017;151(1):23–33.
- Chu DC, Mehta AB, Walkey AJ. Practice patterns and outcomes associated with procalcitonin use in critically ill patients with sepsis. Clin Infect Dis. 2017;64(11):1509–1515.
- Broyles MR. Impact of procalcitonin-guided antibiotic management on antibiotic exposure and outcomes: real-world evidence. open forum infect dis. Fall. 2017;4(4):ofx213.
- Walsh TL, DiSilvio BE, Hammer C, et al. Impact of procalcitonin guidance with an educational program on management of adults hospitalized with pneumonia. Am J Med. 2018;131(2):201 e1–201 e8.
- Nolin TD. Vancomycin and the risk of AKI: now clearer than Mississippi Mud. Clin J Am Soc Nephrol. 2016;11(12):2101–2103.